On-line monitoring system for insulation state of power equipment

Abstract

The application relates to the technical field of power equipment monitoring, and discloses an online monitoring system for the insulation state of power equipment. The system comprises a multi-source data fusion module, which collects running parameters, environmental sensing and historical insulation deterioration record data streams in real time; a spatial insulation field strength distribution module, which combines equipment structure topological maps and an insulation deterioration characteristic library to dynamically map the running parameter data streams to field strength, and generates spatial insulation field strength distribution characteristic information. A reference state topological module constructs an insulation state reference topological network through dynamic weight distribution according to the element service length distribution and the running mode; an insulation deterioration dynamic detection module analyzes the regional field strength deviation of the two, and generates a primary insulation deterioration index. When the primary index exceeds a threshold value, a delay response verification module performs continuous time sequence deterioration trajectory tracking verification; an insulation state trend prediction module simulates multi-step state evolution to obtain predicted field strength distribution characteristics, and a monitoring compensation optimization module corrects the primary index according to the predicted field strength distribution characteristics, and generates an optimized insulation deterioration index.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power equipment monitoring, in particular to an online monitoring system for the insulation state of power equipment. BACKGROUND

[0002] In the process of power system operation, the insulation state of power equipment is directly related to the stability and safety of the entire system. Once the insulation performance deteriorates and is not discovered in time, it is easy to cause equipment failure or even large-scale power outage accidents. At present, the monitoring methods for the insulation state of power equipment mainly include offline monitoring and online monitoring. Offline monitoring requires the equipment to be taken out of operation for detection, which not only affects the normal power supply of the power system, but also has the limitations of long detection period and inability to capture insulation state changes in real time. Especially for key equipment that bears important power supply tasks and cannot be frequently taken out of operation, the applicability of offline monitoring is greatly reduced.

[0003] Although online monitoring technology can realize the collection of insulation parameters of equipment in the running state, the existing online monitoring systems generally have the problem of single data collection dimension. Most of them only focus on the running parameters of the equipment itself (such as partial discharge quantity, dielectric loss value, etc.), and ignore the influence of environmental factors (such as temperature, humidity, and contamination level) on the insulation state, resulting in that the collected data cannot fully reflect the real situation of the equipment insulation. At the same time, when analyzing and processing the collected data, the existing systems lack effective spatial field strength distribution modeling capability, and are difficult to construct insulation state evaluation models that fit the actual operation scene according to the structure topology characteristics of the equipment. Therefore, they can only make qualitative judgments on insulation deterioration, and cannot accurately locate the deterioration area and the deterioration degree.

[0004] The existing online monitoring systems mostly use fixed threshold mode to set the baseline state, without considering the influence of the service time difference of equipment components and the change of operation mode on the insulation baseline state. With the increase of equipment operation time, the insulation performance of components will naturally decay, and the heat generation, electric field distribution, etc. of equipment under different operation modes (such as full load, light load, start-up, shutdown, etc.) also have significant differences. The fixed threshold mode is easy to cause misjudgment or omission. In terms of insulation deterioration trend prediction, the existing systems mostly make simple linear prediction based on historical data, which cannot simulate the dynamic evolution process of insulation state under the joint action of multiple factors, and the accuracy and foresight of the prediction results are insufficient. At the same time, when the system detects suspected insulation deterioration signals, it lacks effective delay response verification mechanism, and it is difficult to distinguish between real deterioration signals and interference signals, further reducing the reliability of the monitoring results. The existence of these problems makes the existing power equipment insulation state monitoring technology difficult to meet the needs of the development of power system towards high reliability and high intelligence, and there is a need for an insulation state online monitoring system that can realize multi-source data fusion, accurate modeling, dynamic baseline setting, reliable detection and accurate prediction. SUMMARY

[0005] The present application aims to provide an online monitoring system for insulation state of power equipment to solve the problems presented in the background.

[0006] To achieve the above-mentioned purpose, the present application provides an online monitoring system for insulation state of power equipment, which comprises:

[0007] A multi-source data fusion module is configured to collect real-time operation parameter data stream, environmental sensing data stream and historical insulation deterioration record data stream of the power equipment.

[0008] A spatial insulation field strength distribution module is configured to perform field strength dynamic mapping on the operation parameter data stream based on the structural topological map of the power equipment and in combination with a preset insulation deterioration feature library, and generate spatial insulation field strength distribution feature information.

[0009] A reference state topology module is configured to construct an insulation state reference topology network through a dynamic weight distribution mechanism according to the element service length distribution and equipment operation mode of the power equipment.

[0010] An insulation deterioration dynamic detection module is configured to perform regional field strength deviation analysis based on the output of the spatial insulation field strength distribution feature information and the insulation state reference topology network, and generate a primary insulation deterioration index.

[0011] A delay response verification module is configured to perform continuous time series deterioration trajectory tracking verification rules when the primary insulation deterioration index exceeds a preset deviation threshold.

[0012] An insulation state trend prediction module is configured to perform multi-step state evolution simulation on the spatial insulation field strength distribution feature information and the environmental sensing data stream, and generate predicted insulation field strength distribution feature information.

[0013] A monitoring compensation optimization module is configured to perform field strength compensation correction on the primary insulation deterioration index based on the predicted insulation field strength distribution feature information, and generate an optimized insulation deterioration index.

[0014] Preferably, the multi-source data fusion module is configured to collect real-time operation parameter data stream, environmental sensing data stream and historical insulation deterioration record data stream of the power equipment, comprising:

[0015] Obtaining the equipment model identification and installation position coordinates of the power equipment;

[0016] Loading the insulation parameter collection protocol set of the corresponding model according to the equipment model identification;

[0017] Activating the environmental sensing array collection node based on the installation position coordinates;

[0018] Synchronously call a historical insulation deterioration feature library index associated with the equipment model identifier;

[0019] Performing a multi-channel concurrent acquisition task generates the operation parameter data stream, the environmental sensing data stream, and the historical insulation deterioration record data stream.

[0020] Preferably, the reference state topology module is configured to construct an insulation state reference topology network through a dynamic weight distribution mechanism according to the element service length distribution of the power equipment and the equipment operation mode, including:

[0021] Analyzing the element service length distribution to generate a service length discrete feature vector;

[0022] Extracting a mode feature code corresponding to the equipment operation mode;

[0023] Establishing an initial topology graph with power equipment elements as vertices and insulation correlations between elements as edges;

[0024] Calculating an initial weight factor for each vertex according to the service length discrete feature vector;

[0025] Fusing the mode feature code to dynamically calibrate the edge connection strength;

[0026] Outputting the insulation state reference topology network with a weight attribute.

[0027] Preferably, the insulation deterioration dynamic detection module is configured to perform regional field strength deviation analysis based on the spatial insulation field strength distribution feature information and the output of the insulation state reference topology network to generate a primary insulation deterioration index, including:

[0028] Segmenting the spatial insulation field strength distribution feature information into field strength partition data sets according to equipment regions;

[0029] Matching the reference field strength intervals of each region in the insulation state reference topology network;

[0030] Calculating the multi-dimensional deviation of the real-time field strength value of each partition relative to the reference field strength interval;

[0031] Aggregating the multi-dimensional deviation of each partition to generate a regional comprehensive deterioration coefficient;

[0032] Converting the regional comprehensive deterioration coefficient into the primary insulation deterioration index based on a preset deterioration level mapping table.

[0033] Preferably, the delay response verification module is configured to perform a continuous time series deterioration trajectory tracking verification rule when the primary insulation deterioration index exceeds a preset deviation threshold, including:

[0034] Recording the first time when the primary insulation deterioration index exceeds the limit as a starting verification timestamp;

[0035] constructing a sliding time window with reference to the starting check timestamp;

[0036] continuously capturing the insulation deterioration index change trajectory within the sliding time window;

[0037] triggering an insulation anomaly confirmation signal when the deterioration trajectory meets the duration threshold and the deterioration monotonicity condition;

[0038] otherwise, resetting the sliding time window and continuously monitoring.

[0039] Preferably, the insulation state trend prediction module is configured to perform multi-step state evolution simulation on the spatial insulation field strength distribution characteristic information and the environmental sensing data stream to generate predicted insulation field strength distribution characteristic information, including:

[0040] extracting the time sequence change pattern of historical insulation field strength distribution characteristic information;

[0041] associating the temperature and humidity change gradient in the environmental sensing data stream;

[0042] constructing a field strength-environment coupling evolution model;

[0043] simulating future multi-step insulation field strength distribution through a rolling time domain prediction algorithm;

[0044] outputting the predicted insulation field strength distribution characteristic information with a confidence interval.

[0045] Preferably, the monitoring compensation optimization module is configured to perform field strength compensation correction on the primary insulation deterioration index based on the predicted insulation field strength distribution characteristic information to generate an optimized insulation deterioration index, including:

[0046] comparing the deviation amount of real-time spatial insulation field strength distribution characteristic information and predicted insulation field strength distribution characteristic information;

[0047] calculating a field strength distribution compensation coefficient when the deviation amount exceeds a tolerance threshold;

[0048] linearly correcting the primary insulation deterioration index according to the field strength distribution compensation coefficient;

[0049] outputting the optimized insulation deterioration index after compensation calibration.

[0050] Preferably, the system further comprises a reference topology adaptive module configured to perform dynamic node weight redistribution on the insulation state reference topology network according to the optimized insulation deterioration index, including:

[0051] analyzing the optimized insulation deterioration index to obtain the deterioration severity level of each region;

[0052] According to the deterioration severity grade, a weight attenuation factor of the corresponding topological node is calculated;

[0053] Based on the weight attenuation factor, a vertex weight of the insulation state reference topological network is updated;

[0054] The mapping relationship between the edge connection strength and the vertex weight is reconstructed.

[0055] Preferably, the system further comprises an insulation field strength reconstruction module, configured to reconstruct an equipment insulation field strength atlas based on the compensated spatial insulation field strength distribution characteristic information when the optimization insulation deterioration index triggers a reconstruction condition, including:

[0056] When the optimization insulation deterioration index continuously exceeds the limit and the compensation correction is invalid, a field strength atlas reconstruction instruction is activated;

[0057] The compensated spatial insulation field strength distribution characteristic information at the current time is obtained;

[0058] The insulation field strength spatial interpolation calculation is performed by fusing the equipment structure topological atlas;

[0059] A new construction equipment insulation field strength atlas with three-dimensional coordinate attributes is generated.

[0060] Preferably, the system further comprises an abnormality disposal decision module, configured to generate an insulation state maintenance strategy according to the reconstructed equipment insulation field strength atlas, including:

[0061] The field strength weak area coordinates in the new construction equipment insulation field strength atlas are identified;

[0062] The operation and maintenance procedure library of the power equipment is associated to match a disposal plan template;

[0063] In combination with the real-time load rate of the equipment, a disposal priority sequence is generated;

[0064] An insulation state maintenance strategy containing a disposal location coordinate, a disposal mode and an execution time sequence is output.

[0065] Compared with the prior art, the beneficial effects of the present application are:

[0066] The operation parameter data stream, the environmental sensing data stream and the historical insulation deterioration record data stream of the power equipment are collected and integrated in real time by the multi-source data fusion module, breaking the limitation of single data collection dimension of the existing monitoring system, and being able to comprehensively capture various factors affecting the insulation state of the equipment, ensuring that the collected data are more in line with the real operation scene of the equipment insulation, and providing more rich and comprehensive basic data support for subsequent insulation state analysis. Compared with the traditional system which only focuses on the operation parameters of the equipment itself, the present system considers the influence of environmental factors and historical deterioration records, so that the evaluation of the insulation state is no longer limited to a single dimension, but reflects the insulation condition of the equipment from multiple dimensions and multiple angles, effectively avoiding the evaluation deviation caused by the one-sidedness of the data.

[0067] The spatial insulation field strength distribution module dynamically maps the operation parameter data stream based on the equipment structure topology map and in combination with the preset insulation deterioration feature library, generates spatial insulation field strength distribution feature information, and solves the problem of lack of spatial field strength modeling capability in the existing system. By constructing a spatial field strength model matched with the actual structure of the equipment, the electric field distribution in different regions inside the equipment can be accurately presented, not only the qualitative judgment of insulation deterioration can be realized, but also the deterioration region can be accurately located, providing a clear direction for subsequent maintenance and repair, avoiding the disadvantages of the traditional system that cannot locate the deterioration region and the blindness of maintenance work. At the same time, the introduction of the preset insulation deterioration feature library enables the field strength dynamic mapping process to combine the known deterioration rules, further improving the accuracy and pertinence of the spatial insulation field strength distribution feature information.

[0068] The reference state topology module constructs an insulation state reference topology network through a dynamic weight distribution mechanism according to the service length distribution of the equipment elements and the equipment operation mode, changing the mode of setting the reference state by using a fixed threshold in the existing system. This module fully considers the natural insulation performance decay caused by the service length difference of the equipment elements, and the influence of the working condition change of the equipment under different operation modes on the insulation reference state, and through dynamic weight distribution, the reference state can be adaptively adjusted with the change of the equipment running time and the operation mode, ensuring that the reference state always fits the actual operation of the equipment. This dynamic reference setting mode effectively avoids the misjudgment and omission problem caused by equipment aging or operation mode change in the fixed threshold mode, improving the accuracy of the insulation state evaluation.

[0069] The insulation deterioration dynamic detection module performs regional field intensity deviation degree analysis based on the spatial insulation field intensity distribution characteristic information and the output of the insulation state benchmark topology network, generates a primary insulation deterioration index, and realizes quantitative analysis of insulation deterioration. Through regional field intensity deviation degree analysis, the difference degree of the insulation state of each region and the benchmark state can be accurately calculated, and the severity of the deterioration is clear. Compared with the qualitative judgment of the traditional system, it is more scientific and practical. When the primary insulation deterioration index exceeds the preset deviation threshold, the delay response verification module executes continuous time series deterioration trajectory tracking verification rules, which can effectively distinguish between real deterioration signals and interference signals. Through continuous time series tracking, the change trend of the deterioration index is observed, false alarms caused by instantaneous interference signals are avoided, the reliability of the monitoring results is improved, and unnecessary maintenance costs and workload are reduced.

[0070] The insulation state trend prediction module performs multi-step state evolution simulation on the spatial insulation field intensity distribution characteristic information and the environmental sensing data stream, generates predicted insulation field intensity distribution characteristic information, and breaks through the limitations of the simple linear prediction of the traditional system. Multi-step state evolution simulation can comprehensively consider the dynamic changes of equipment operating parameters and environmental factors, simulate the development trend of the insulation state under different conditions, make the prediction results more forward-looking and accurate, help operation and maintenance personnel to grasp the possible direction and degree of insulation deterioration in advance, and provide effective reference for formulating preventive maintenance strategies. The monitoring compensation optimization module compensates and corrects the primary insulation deterioration index based on the predicted insulation field intensity distribution characteristic information, generates an optimized insulation deterioration index, and further improves the accuracy of the deterioration index. By introducing prediction information for compensation correction, errors caused by the limitations of the current monitoring data or environmental fluctuations can be eliminated, and the optimized insulation deterioration index output at the end can more truly reflect the actual deterioration condition of the equipment insulation, providing a more reliable basis for subsequent operation and maintenance decisions. BRIEF DESCRIPTION OF DRAWINGS

[0071] Figure 1 A timing diagram of the power equipment insulation state online monitoring system described in the present application;

[0072] Figure 2 A flowchart of the benchmark state topology module;

[0073] Figure 3 A flowchart of the insulation deterioration dynamic detection module;

[0074] Figure 4 A flowchart of the insulation state trend prediction module;

[0075] Figure 5 A flowchart of the benchmark topology adaptive module. DETAILED DESCRIPTION

[0076] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0077] With reference to Figure 1 The present application provides an online monitoring system for insulation state of power equipment, which comprises:

[0078] The multi-source data fusion module collects the running parameter data stream, the environmental sensing data stream and the historical insulation degradation record data stream of the power equipment in real time. The spatial insulation field strength distribution module dynamically maps the running parameter data stream based on the structural topological map of the power equipment and in combination with the preset insulation degradation feature library, to generate spatial insulation field strength distribution feature information. The reference state topological module constructs an insulation state reference topological network through a dynamic weight distribution mechanism according to the element service time distribution and the equipment operation mode of the power equipment. The insulation degradation dynamic detection module performs regional field strength deviation analysis based on the output of the spatial insulation field strength distribution feature information and the insulation state reference topological network, to generate a primary insulation degradation index. The delay response verification module executes continuous time series degradation trajectory tracking verification rules when the primary insulation degradation index exceeds a preset deviation threshold. The insulation state trend prediction module performs multi-step state evolution simulation on the spatial insulation field strength distribution feature information and the environmental sensing data stream, to generate predicted insulation field strength distribution feature information. The monitoring compensation optimization module performs field strength compensation correction on the primary insulation degradation index based on the predicted insulation field strength distribution feature information, to generate an optimized insulation degradation index.

[0079] Embodiment 1: With reference to Figure 2 The start of the multi-source data fusion module is the identification of the target power equipment. The system first acquires the unique equipment model identification of the equipment, which is usually provided by the equipment nameplate information or the system registration database, and contains the key information of the equipment, such as the manufacturer, the specification and model, and the rated parameters. The installation position coordinates of the equipment are acquired, which are in the geographic information system standard format and accurate to the meter level, and are used to locate the physical position of the equipment. Based on the equipment model identification, the system loads the insulation parameter acquisition protocol set of the corresponding model from the preconfigured protocol library. The protocol set specifies the parameters of data acquisition in detail: for the running parameters, the acquisition frequency is set to 10 times per second, the measurement accuracy is required to reach 0.5 level, and the data format adopts the IEEE floating point number standard; for the environmental parameters, the acquisition interval is set to 30 seconds, the temperature measurement range covers-40℃ to +85℃, and the humidity measurement accuracy is required to be ±3%RH. The protocol set also contains data verification rules, such as the generation and verification method of the CRC cyclic redundancy check code.

[0080] The installation position coordinates are used to activate the environmental sensing array collection nodes deployed around the equipment. These nodes are distributed according to a pre-planned topology, including temperature sensors installed on the surface of the transformer oil tank, humidity sensors arranged near the radiator, and air quality monitoring points set around the equipment. Each collection node has an independent network address, and the system establishes a communication connection with each node through the Modbus TCP protocol. At the same time of starting real-time data collection, the system synchronously calls the historical insulation degradation feature library index associated with the equipment model identification. The index is organized using a B+ tree data structure, supporting fast retrieval. The index content includes all insulation-related events recorded in the past operation of the model equipment, such as partial discharge historical data, medium loss factor change records, insulation oil chromatographic analysis results, etc. These historical data are sorted by timestamp and stored in a large-capacity time series database.

[0081] When performing multi-channel concurrent collection tasks, the system uses a multi-thread architecture. The running parameter collection thread obtains real-time running data through the monitoring IED (intelligent electronic device) of the equipment, including the effective value and phase angle of three-phase voltage and current, the instantaneous value of active power and reactive power, and the reading of winding temperature meter. The environmental sensing data collection thread simultaneously obtains environmental parameters such as temperature, humidity, and air pressure from multiple sensor nodes. The historical data retrieval thread extracts insulation degradation-related records from the historical database for the past three years in parallel. All data streams are marked with a unified timestamp, with a time synchronization accuracy of milliseconds, ensuring the consistency of data from different sources in the time dimension. The collected running parameter data stream contains electrical measurement values of the equipment, and these data are processed by digital filtering to eliminate noise interference. The environmental sensing data stream contains not only basic temperature and humidity readings, but also sensor state information such as battery power and communication signal strength. The historical insulation degradation record data stream provides trend data of insulation performance changes in long-term operation of the equipment. The three types of data streams are cached in different memory buffers, waiting for subsequent processing module calls.

[0082] The module first parses the element service length distribution of the power equipment, which comes from the equipment account management system. For equipment such as transformers, the elements include high-voltage windings, low-voltage windings, cores, insulating oil, bushings, tap changers, etc. The operation time of each element is accurately recorded, and the system calculates the cumulative operating hours since operation, generating a service length discrete feature vector. The vector is normalized to map the actual running time to the interval [0, 1], where 0 represents a new state and 1 represents reaching the design life. The identification of the equipment operation mode is based on the real-time operation parameter data stream, and the system continuously monitors the load rate, operating temperature, cooling system state and other parameters of the equipment. According to the preset mode classification rules, the equipment operating state is divided into multiple modes: light load operation mode, rated operation mode, overload operation mode, and no-load operation mode. Each mode corresponds to a unique mode feature code, which is represented by one-hot encoding for easy computer processing.

[0083] When establishing the initial topology graph, each element of the power equipment is taken as a vertex, and each vertex contains the identifier, type attribute, and physical location information of the element. The insulation correlation between elements is taken as an edge, which represents the mutual influence relationship on insulation. The initial value of the connection strength of the edge is set according to the physical distance between elements and the insulation material characteristics. The negative correlation algorithm is used in the calculation, that is, the longer the service time of the element, the smaller the weight factor value, indicating that the influence weight of the element on the overall insulation state should be reduced. The specific value of the weight factor is calculated by a linear decay function, considering the importance coefficient of the element type.

[0084] When the equipment is in an overload operation mode, the connection strength of the insulation correlation edge related to heat generation will increase; when the environmental humidity is high, the connection strength of the edge related to surface insulation will be adjusted. The calibration process uses a weighted average algorithm, and the weight coefficient of the real-time operation mode is higher than the historical statistical value. The final output of the insulation state benchmark topology network is a weighted directed graph structure. Each vertex has a corresponding weight factor, and each edge has a dynamically adjusted connection strength value. This network serves as a benchmark model for insulation state assessment, and truly reflects the expected insulation performance level of the equipment under the current operating state and service length conditions. The benchmark topology network will be continuously updated as the equipment operating time and operating state change, ensuring that the benchmark model always keeps pace with the actual condition of the equipment.

[0085] Example 2: see Figure 3The processing procedure of the insulation deterioration dynamic detection module starts from the regional segmentation of the spatial insulation field intensity distribution characteristic information. The high-voltage winding region is further subdivided into upper, middle and lower parts, and the low-voltage winding region adopts the same subdivision manner. The core and clamp regions are divided into A, B and C three-phase partitions according to the phase. The insulation oil region is divided into multiple flow partitions according to the oil channel structure. This fine regional division ensures the accuracy of the field intensity analysis. In the generation process of the field intensity partition data set, each partition contains a group of field intensity sampling points. For the winding region, the sampling points are uniformly distributed along the winding height direction; for the insulation oil region, the sampling points are set at the main oil channel positions; for the bushing region, the sampling points are distributed along the axial and radial directions of the bushing. Each sampling point records its three-dimensional coordinate information and real-time field intensity measurement value, and these data come from the field intensity sensor array installed inside the device. The original data collected by the sensor are processed through signal conditioning and digitization, and are converted into standardized field intensity values.

[0086] When matching the reference field intensity interval of each region in the reference topology network of the insulation state, the system calls the pre-established reference model. The model is obtained based on the insulation performance data and historical normal operation data at the initial stage of the device operation. Each partition corresponds to a reference field intensity interval, which is defined by an upper limit value and a lower limit value. When calculating the multi-dimensional deviation degree of the real-time field intensity value of each partition relative to the reference field intensity interval, the system uses multiple deviation degree indexes. The absolute value deviation degree calculates the absolute deviation of the real-time field intensity from the median value in the reference interval; the gradient deviation degree analyzes the rate of field intensity change, which is obtained by comparing the difference between the current field intensity value and the historical data; the distribution form deviation degree evaluates the uniformity of the field intensity distribution in the partition, and calculates the dispersion degree of the field intensity value by using statistical methods. Each deviation degree is normalized to convert it into a dimensionless value in the range of [0, 1].

[0087] The process of aggregating the multi-dimensional deviation degree of each partition to generate the regional comprehensive deterioration coefficient uses a weighted fusion algorithm. The weights of different partitions are allocated according to their importance in the insulation system, and the weight of the winding region is higher than that of other regions, and the weight of the high-voltage region is higher than that of the low-voltage region. The weight of each deviation degree index is also adjusted according to its sensitivity, and the weight of the gradient deviation degree is usually higher than that of other indexes. The aggregation calculation uses linear weighted summation to produce a deterioration coefficient value that comprehensively reflects the insulation state of the region. When converting the regional comprehensive deterioration coefficient into the primary insulation deterioration index based on the preset deterioration level mapping table, the system calls the pre-defined mapping rule. The mapping table discretizes the continuous deterioration coefficient value into several insulation state levels, for example: 0-0.2 corresponds to the normal state, 0.2-0.4 corresponds to the attention state, 0.4-0.6 corresponds to the abnormal state, and 0.6 and above corresponds to the dangerous state. Each state level is equipped with detailed state description and processing suggestion. The converted primary insulation deterioration index contains region identification, state level and time stamp information.

[0088] A sliding time window is constructed with the starting check time stamp as the reference, and the window size is dynamically adjusted according to the device type and operating state. The default window size for transformer-type devices is 30 minutes, and the window size can be set to a fixed or variable length. The variable window is adaptively adjusted according to the change in the degradation index, and the window size is reduced when the index changes rapidly and expanded when the index is stable. Within the window, the system continuously captures the trajectory of the insulation degradation index, which includes the degradation index value, the change rate, and the acceleration. The collection frequency is adjusted according to the degree of degradation, and the initial collection frequency is once per second, which is increased when the index rises. The trajectory data is stored in a ring buffer for easy access and analysis. When the degradation trajectory meets the continuous time threshold and the monotonicity condition of degradation, the system triggers an insulation abnormality confirmation signal. The continuous time threshold requires that the index exceed the limit for a certain time within the window, such as 5 consecutive minutes of over-limit state. The monotonicity of degradation requires that the index generally rises, allowing small fluctuations, which is determined by calculating the linear regression slope through trend analysis. If the conditions are not met, the system resets the window and continues monitoring, which includes clearing the buffer, resetting the time stamp, adjusting the parameters, and other operations. During monitoring, the system records changes in the environment and device operating state for subsequent analysis and diagnosis.

[0089] The output of the insulation degradation dynamic detection module not only contains simple state identification, but also provides detailed analysis data: the deviation component of each region, the aggregation weight distribution, the history record of state transition, etc. These additional information provides rich data support for subsequent trend prediction and maintenance decision. The verification result of the delayed response verification module also contains detailed trajectory analysis report, including time window parameters, trajectory feature description, verification decision basis, etc. The parameter setting of the system supports online adjustment, and users can modify various thresholds and parameters according to the characteristics of the device and operating experience.

[0090] Embodiment 3: refer to Figure 4 The processing process of the insulation state trend prediction module starts with the extraction of the time sequence change pattern of the historical insulation field strength distribution characteristics information. The module accesses the historical database stored by the system to obtain the field strength distribution data of each region of the device in the past period of time. These data are organized in time sequence, including the field strength measurement value and its spatial distribution information at each sampling time point. The pattern extraction uses time series analysis method to identify the periodic changes, trend changes and random fluctuations in the field strength data. For devices such as transformers, the field strength changes often show daily cycle characteristics related to load changes and long cycle characteristics related to seasonal changes. The system uses sliding window technology to analyze recent data, and the window size is adaptively adjusted according to the device type and operating history. During pattern recognition, the system will pay special attention to the mutation points of field strength change, which may correspond to significant changes in device operating state or sudden deterioration of insulation conditions.

[0091] Correlating temperature and humidity change gradients in environmental sensor data streams is a key step in trend prediction. Environmental data comes from a multi-sensor network deployed around the device, including temperature sensors, humidity sensors, barometric pressure sensors, etc. Temperature change gradients are obtained by calculating the difference between adjacent time point temperature values, and humidity change gradients are calculated in a similar manner. These gradient values reflect the dynamic change characteristics of environmental conditions, which are closely related to the performance changes of insulating materials. For example, temperature rise will accelerate the aging process of insulating materials, and humidity increase may reduce surface insulation strength. The system establishes a correlation model between environmental parameters and insulation field strength, which is based on statistical analysis of historical data to quantify the typical change pattern of field strength under different environmental conditions.

[0092] Building a field strength-environment coupling evolution model is the core step of trend prediction, which links field strength changes with environmental factor changes to form a multivariate prediction framework. The model considers the time series correlation of field strength itself and the lagging effect of environmental factors on field strength. Model parameters are trained from historical data by machine learning methods and updated regularly to maintain prediction accuracy. The mathematical expression of the model is:

[0093]

[0094] Where: represents the future time field strength prediction value, represents the current and historical field strength observation value, and represent the current temperature and humidity measurement values, and represent the current temperature and humidity change gradients, represents the model parameter set. This nonlinear function relationship is realized through a neural network structure, which can capture the complex interaction between field strength and environmental factors.

[0095] Simulate future multi-step insulation field strength distribution through rolling time domain prediction algorithm. This algorithm uses an iterative prediction method, first predicting the field strength distribution at the next time point, then using the predicted value as the new observation value to continue predicting the state at a further time. This rolling prediction method can continuously incorporate the latest observation data, reducing the accumulation of prediction errors. During the multi-step prediction process, the system simultaneously predicts the trend of environmental parameter changes as input conditions for field strength prediction. The prediction step length is set according to application requirements, usually including short-term prediction (a few minutes to a few hours) and medium-term prediction (several hours to several days). Each prediction step generates complete field strength spatial distribution information, including predicted values for all monitoring regions.

[0096] The prediction insulation field strength distribution characteristic information with confidence interval is output, and the confidence interval is obtained by analyzing the statistical characteristics of historical prediction errors, reflecting the uncertainty range of the prediction results. The system provides upper and lower limit estimates for each prediction value to form a prediction interval, and the confidence level can be set as needed, usually 95%. The prediction results are output in a standardized format, including fields such as timestamp, prediction value, and confidence interval boundaries, and are stored in a prediction database for subsequent module calls for analysis. The monitoring compensation optimization module is implemented based on the comparison and analysis of the predicted insulation field strength distribution characteristic information and real-time monitoring data. The module first compares the deviation between real-time and predicted insulation field strength distribution characteristic information, calculates the absolute and relative differences between the real-time measurement value and the prediction value at each monitoring point using point-by-point comparison, records the spatial distribution and temporal variation characteristics of the deviation, analyzes the deviation rules and abnormalities, and calculates the field strength distribution compensation coefficient when the deviation exceeds the threshold. The threshold is dynamically adjusted according to the device type, operating state, and historical deviation statistics, and is usually set to the upper percentile value of the historical deviation distribution. The compensation coefficient calculation considers the deviation size, duration, and spatial consistency, and local deviation is adjusted using local compensation strategy, while global deviation is adjusted using system-level compensation. The compensation coefficient is a multi-dimensional vector corresponding to the compensation amount in different regions and time points.

[0097] The primary insulation deterioration index is linearly corrected according to the field strength distribution compensation coefficient, and the correction process maps the compensation coefficient into the calculation formula of the deterioration index to adjust the index value to reflect the prediction information. The correction weight is dynamically allocated according to the reliability of the prediction results, with higher weight given when the prediction confidence interval is narrower and lower weight given when the confidence interval is wider. The corrected deterioration index more accurately reflects the actual trend of the insulation state. The optimized insulation deterioration index after compensation calibration is output. This index contains basic state level information and detailed compensation adjustment records. The system records metadata such as time, compensation amount, and correction reason for each compensation operation to form a complete compensation log. The optimized index value is used for subsequent alarm judgment and maintenance decision-making, and is also fed back to the prediction model to improve future prediction accuracy.

[0098] The data storage and management adopts a hierarchical architecture, with raw monitoring data, prediction data, and compensation data stored in different database layers. The data access interface provides unified data retrieval services to support the data needs of various analysis applications. The system performance monitoring module continuously tracks the processing delay and resource consumption of prediction and compensation calculations to ensure that the system meets the real-time requirements of online monitoring. The collaborative work of the trend prediction module and the compensation optimization module forms a closed-loop optimization system. The prediction results guide the compensation adjustment, and the monitoring data after compensation is used to improve the prediction model, and this iterative optimization mechanism enables the system to continuously adapt to changes in device state, improving the accuracy and reliability of monitoring results. The system also provides a manual intervention interface to allow experienced engineers to review and adjust the automatic prediction and compensation results, combining artificial intelligence analysis with human expert experience.

[0099] Example 4: refer to Figure 5 The processing procedure of the reference topology adaptive module is based on the analysis results of the optimization insulation deterioration index. The module receives the data stream of the optimization insulation deterioration index from the monitoring compensation optimization module, which is organized by device area and contains the identifier of each area, the value of the deterioration index, the timestamp, and the confidence information. The analysis procedure first verifies the validity of the index data, checks the data integrity and time consistency, and then processes it by area grouping. The system obtains the deterioration severity level of each area by a pre-defined level division rule, which maps the continuous deterioration index value to discrete severity levels. For example, the index value between 0.0 and 0.2 corresponds to level 0 (normal state), 0.2 to 0.4 corresponds to level 1 (slight deterioration), 0.4 to 0.6 corresponds to level 2 (moderate deterioration), 0.6 to 0.8 corresponds to level 3 (severe deterioration), and above 0.8 corresponds to level 4 (dangerous state). Each level has corresponding color coding and description text for visual display and subsequent processing.

[0100] When calculating the weight attenuation factor of the corresponding topology node according to the deterioration severity level, the system uses a level mapping algorithm. The algorithm assigns an attenuation coefficient to each deterioration level, and the higher the level, the larger the attenuation coefficient, indicating that the weight of the node in the reference topology network should be correspondingly reduced. In the calculation process, the historical state changes of the node are considered, and if the node state continuously deteriorates, a larger attenuation step is used; if the state improves, the attenuation degree is appropriately reduced. The calculation of the attenuation factor also refers to the initial weight of the node and the state of the adjacent nodes to maintain the overall consistency of the network structure.

[0101] When updating the vertex weight of the insulation state reference topology network based on the weight attenuation factor, the system uses a gradual adjustment strategy. The update process is not a complete reset of the weight, but an application of the attenuation factor based on the original weight, making the weight change smooth transition. The weight update record of each vertex includes the timestamp, the original weight value, the new weight value, and the adjustment reason. The weight update algorithm ensures that the total weight of the network remains unchanged, only the weight proportion of each node is redistributed. The system also records the historical trajectory of weight adjustment for analyzing the evolution trend of the network structure. The edge connection strength reflects the insulation correlation degree between nodes, and when the vertex weight changes, the edge strength connected to it also needs to be adjusted accordingly. The system establishes a functional relationship between edge strength and vertex weight, and when the vertex weight attenuates, the connection strength of the output edge is correspondingly weakened, and the connection strength of the input edge is adjusted according to the state of the adjacent nodes. The reconstruction of the mapping relationship uses an iterative optimization algorithm to ensure the stability and consistency of the network structure. The completed insulation state reference topology network will serve as the reference for the next round of insulation state evaluation, realizing the dynamic adaptation and continuous optimization of the evaluation standard.

[0102] When the insulation deterioration index continues to exceed the limit and the compensation correction is ineffective, the system activates the field strength map reconstruction instruction. The judgment standard for continuous over-limit is that the deterioration index exceeds the threshold for multiple consecutive monitoring periods, and the compensation operation fails to make the index fall within the normal range. The determination of ineffective compensation correction is based on the trend of the index after compensation. If the index value continues to deteriorate or shows no signs of improvement, it is considered that the compensation correction is ineffective. The reconstruction instruction contains parameters such as reconstruction range, reconstruction accuracy and reconstruction time, which are dynamically set according to the severity of deterioration and the importance of equipment. When obtaining the spatial insulation field strength distribution characteristic information after compensation correction at the current time, the system reads the latest field strength data from the data cache area. These data have been processed by the monitoring compensation optimization module and contain compensation adjustments based on prediction information. The data format includes the three-dimensional coordinates of each monitoring point, field strength measurement value, compensation amount and time stamp. The system performs quality checks on the data, eliminating outliers and unreliable data, and then performs data alignment and format conversion to prepare for subsequent reconstruction calculations.

[0103] Fusion equipment structure topology map for insulation field strength spatial interpolation calculation is the core link of the reconstruction process, and the equipment structure topology map provides the geometric structure and material distribution information of the equipment, including winding position, insulation barrier, cooling channel and other structural characteristics. Spatial interpolation calculation uses an interpolation algorithm based on a physical model, which considers the dielectric properties of insulation materials, electric field distribution rules and equipment structure characteristics. The interpolation process first discretizes the equipment space into a three-dimensional grid, and then calculates the field strength value of the grid node based on the known monitoring point data by solving the electric field distribution equation. The interpolation algorithm uses adaptive grid refinement technology, using finer grids in areas with rapid field strength changes and coarser grids in areas with uniform field strength, balancing calculation accuracy and efficiency.

[0104] When generating a new insulation field strength map for the constructed equipment with three-dimensional coordinate attributes, the system combines the interpolation calculation results with the equipment coordinate system. Each grid node contains three-dimensional coordinate information and the calculated field strength value, forming a complete field strength spatial distribution model. The map data is organized in a hierarchical structure, including the basic geometry layer, field strength data layer, material attribute layer and annotation information layer. The system adds metadata information to the map, including generation time, data version, calculation parameters and accuracy indicators. The version management of the new map and the old map uses incremental update, recording the change content and impact range of each reconstruction. Referring to Table 1, the correspondence between the optimized insulation deterioration index and the weight decay factor is shown.

[0105] Table 1: Correspondence between deterioration level and weight decay factor

[0106]

[0107] The system performance monitoring module tracks the computational load of the reference topology adaptation and field strength reconstruction in real-time, dynamically adjusting the allocation of computational resources. When the device state changes rapidly, the system automatically increases the computational priority and speeds up the update of the network and the map; when the state is stable, the computational frequency is appropriately reduced to save system resources. The resource management strategy ensures that the system can complete the reference adjustment and map reconstruction tasks in a timely manner under various operating conditions. The error handling and recovery mechanism ensures the reliability of the implementation. When an exception occurs during the computation, the system can automatically save the intermediate results, revert to the previous stable state, and continue execution after the problem is resolved. All abnormal events are recorded in the system log, including the type of exception, the time of occurrence, the processing process, and the final result. The log analysis tool helps operation and maintenance personnel identify system problems and improve the implementation process.

[0108] The visualization module provides an intuitive graphical interface for the implementation results. The adjustment of the reference topology network is displayed in the form of a topology graph, with vertex size representing weight value, edge thickness representing connection strength, and color representing state level. The field strength map uses three-dimensional rendering technology for display, using a color gradient to represent field strength, and supporting interactive operations such as rotation, scaling, and cross-section viewing. The visualization tool helps users understand the implementation effect and decision basis. All data exchanges during the implementation process use standardized data formats and interface protocols. The reference topology network data uses graph structure description language, and the field strength map data uses scientific data format, ensuring data portability and interoperability. The system provides data export functions, supporting the output of implementation results to other analysis tools or management systems for further processing.

[0109] Example 5: Matching the operation and maintenance procedure library of associated power equipment with the disposal plan template is a key step in decision-making. The operation and maintenance procedure library stores standard maintenance procedures provided by equipment manufacturers, industry specification requirements, and experience disposal plans accumulated by operating units. The system retrieves matching disposal plans in the procedure library according to device type, fault mode, and degradation level, etc. The matching process uses a rule-based reasoning algorithm, first determines the applicable procedure subset according to the device model, then filters possible disposal methods according to the degradation type, and finally adjusts the disposal intensity according to the degradation level. Each matched plan template contains detailed information such as disposal operation steps, required tools and materials, safety precautions, estimated man-hours, and acceptance criteria. The system also records the historical execution effect data of each plan as a reference for this matching.

[0110] When generating the treatment priority sequence in combination with the real-time load rate of the equipment, the system comprehensively considers technical factors and operational requirements. Real-time load rate data comes from the equipment monitoring system, reflecting the current operational importance and the impact of power failure. During periods of high load rate, power failure maintenance is generally avoided as much as possible, while during periods of low load rate, maintenance work is more suitable. The priority calculation algorithm combines technical urgency and operational constraints, with technical urgency based on deterioration development speed and risk consequence assessment, and operational constraints including load level, weather conditions, and backup equipment status. The system calculates a comprehensive priority score for each treatment item, with higher scores placed at the front of the execution sequence. The priority sequence also takes into account the dependencies between treatment items, as some treatment operations require completion of other operations, and these logical relationships are modeled as constraints and incorporated into the sorting algorithm.

[0111] The output insulation state maintenance strategy is the final result of the entire implementation process. The strategy document is organized in a structured format, including treatment location coordinates, treatment methods, and execution timing. Treatment location coordinates are described in both equipment coordinate system and geographic coordinate system, containing both relative position in the equipment structure and actual geographic location information, facilitating on-site personnel positioning. Treatment methods describe detailed operation steps and technical requirements, including specific specifications for different processing methods such as cleaning, drying, repainting, and replacement. Execution timing provides suggested implementation time window, estimated duration, and progress schedule, taking into account factors such as equipment operation plan, weather forecast, and resource preparation. The strategy output also includes a risk assessment report, explaining the consequences of not treating and the risks during the treatment process. The system provides multiple output formats, including machine-readable XML format, human-readable PDF document, and API interface for integration with other management systems.

[0112] The implementation of the abnormal treatment decision module fully embodies the combination of professional knowledge and data-driven. The system not only relies on mathematical models and algorithms for analysis, but also deeply integrates domain knowledge and operational experience in equipment insulation management. The decision-making process considers the balance between technical feasibility and operational actual conditions, generating maintenance strategies that ensure technical rationality and on-site operability. The implementation results provide scientific basis and technical support for equipment maintenance management, helping operational units optimize maintenance resource allocation and improve the precision and effectiveness of equipment maintenance. Data security and permission management are fully considered in the implementation process. Sensitive information in the maintenance strategy is protected by access control, and users of different levels can only view content within their respective permission range. Encryption measures are used in the strategy transmission and storage process to prevent information leakage. All access and modification operations on the strategy are recorded in detail, forming a complete audit trail.

[0113] The system also provides policy simulation and impact analysis functions, allowing users to simulate the effects of policy execution before implementation, evaluate the impact of different scheduling, and compare the pros and cons of various treatment options. Simulation results help users make more informed decisions and optimize the overall effectiveness of maintenance plans. Impact analysis can evaluate the comprehensive impact of maintenance work on equipment reliability, operating costs, and system risks, providing a comprehensive reference for decision-making. The implementation of the abnormal treatment decision module completes the complete closed loop from condition monitoring to maintenance action. The system not only can find insulation problems of equipment, but also can provide specific treatment options and implementation suggestions, and convert monitoring data into actual productivity. This complete solution reflects the ultimate value of the condition monitoring system and provides reliable technical support for the safe operation and intelligent maintenance of power equipment. The decision-making data and experience accumulated during the implementation process will also continuously enrich the knowledge base of the system, promoting the continuous evolution and improvement of decision-making capabilities.

[0114] It should be noted that the relational terms herein, such as first and second, are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any such actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0115] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and variations can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended relevant limitations and their equivalents.

Claims

1. An on-line monitoring system for insulation condition of power equipment, characterized in that, The system comprises: A multi-source data fusion module for collecting power equipment operation parameter data stream, environmental sensing data stream and historical insulation deterioration record data stream in real time; A spatial insulation field strength distribution module for dynamically mapping the operation parameter data stream based on the structural topological map of the power equipment and combining a preset insulation deterioration feature library to generate spatial insulation field strength distribution feature information; A reference state topology module for constructing an insulation state reference topology network through a dynamic weight distribution mechanism according to the element service length distribution and equipment operation mode of the power equipment, specifically comprising: Analyzing the element service length distribution to generate a service length discrete feature vector; Extracting the mode feature code corresponding to the equipment operation mode; Establishing an initial topological graph with power equipment elements as vertices and insulation correlation between elements as edges; Calculating the initial weight factor of each vertex according to the service length discrete feature vector; Fusing the mode feature code to dynamically calibrate the edge connection strength; Outputting the insulation state reference topology network with weight attributes; An insulation deterioration dynamic detection module for performing regional field strength deviation analysis based on the output of the spatial insulation field strength distribution feature information and the insulation state reference topology network to generate a primary insulation deterioration index, specifically comprising: Segmenting the spatial insulation field strength distribution feature information into field strength partition data sets according to equipment regions; Matching the reference field strength interval of each region in the insulation state reference topology network; Calculating the multi-dimensional deviation of the real-time field strength value of each partition relative to the reference field strength interval; Aggregating the multi-dimensional deviation of each partition to generate a regional comprehensive deterioration coefficient; Converting the regional comprehensive deterioration coefficient into the primary insulation deterioration index based on a preset deterioration grade mapping table; A delay response verification module for executing continuous time series deterioration trajectory tracking verification rules when the primary insulation deterioration index exceeds a preset deviation threshold, specifically comprising: Recording the first time when the primary insulation deterioration index exceeds the limit as the starting verification time mark; Constructing a sliding time window with the starting verification time mark as the reference; Continuously capturing the insulation deterioration index change trajectory within the sliding time window; Triggering an insulation anomaly confirmation signal when the deterioration trajectory meets the continuous duration threshold and the deterioration monotonicity condition; Otherwise, reset the sliding time window and continue monitoring; An insulation state trend prediction module for performing multi-step state evolution simulation on the spatial insulation field strength distribution feature information and environmental sensing data stream to generate predicted insulation field strength distribution feature information; A monitoring compensation optimization module for performing field strength compensation correction on the primary insulation deterioration index based on the predicted insulation field strength distribution feature information to generate an optimized insulation deterioration index.

2. The on-line monitoring system for insulation condition of power equipment according to claim 1, characterized in that, The multi-source data fusion module is used for collecting power equipment operation parameter data stream, environmental sensing data stream and historical insulation deterioration record data stream in real time, comprising: Obtaining the equipment model identification and installation location coordinates of the power equipment; Loading the insulation parameter collection protocol set corresponding to the equipment model identification according to the equipment model identification; Activating the environmental sensing array collection node based on the installation location coordinates; Synchronously calling the historical insulation deterioration feature library index associated with the equipment model identification; The performing multi-channel concurrent acquisition task generates the running parameter data stream, the environmental sensing data stream, and the historical insulation deterioration record data stream.

3. The on-line monitoring system for insulation condition of power equipment according to claim 1, characterized in that, The insulation state trend prediction module is configured to perform multi-step state evolution simulation on the spatial insulation field intensity distribution characteristic information and the environmental sensing data stream to generate predicted insulation field intensity distribution characteristic information, including: extracting a time sequence variation pattern of historical insulation field intensity distribution characteristic information; associating temperature and humidity variation gradients in the environmental sensing data stream; constructing a field intensity-environment coupling evolution model; simulating future multi-step insulation field intensity distribution through a rolling horizon prediction algorithm; outputting the predicted insulation field intensity distribution characteristic information with a confidence interval.

4. The on-line monitoring system for insulation condition of power equipment according to claim 1, characterized in that, The monitoring compensation optimization module is configured to perform field intensity compensation correction on the primary insulation deterioration index based on the predicted insulation field intensity distribution characteristic information to generate an optimized insulation deterioration index, including: comparing deviation amounts of real-time spatial insulation field intensity distribution characteristic information and predicted insulation field intensity distribution characteristic information; when the deviation amount exceeds a tolerance threshold, calculating a field intensity distribution compensation coefficient; linearly correcting the primary insulation deterioration index according to the field intensity distribution compensation coefficient; outputting the optimized insulation deterioration index after compensation calibration.

5. The on-line monitoring system for insulation condition of power equipment according to claim 1, characterized in that, The system further includes a reference topology adaptive module configured to perform dynamic node weight redistribution on the insulation state reference topology network according to the optimized insulation deterioration index, including: analyzing the optimized insulation deterioration index to obtain deterioration severity levels of each region; calculating weight attenuation factors of corresponding topology nodes according to the deterioration severity levels; updating vertex weights of the insulation state reference topology network based on the weight attenuation factors; reconstructing a mapping relationship between edge connection strengths and vertex weights.

6. The on-line monitoring system for insulation condition of power equipment according to claim 1, characterized in that, The system further includes an insulation field intensity reconstruction module configured to reconstruct a device insulation field intensity map based on compensated spatial insulation field intensity distribution characteristic information when the optimized insulation deterioration index triggers a reconstruction condition, including: when the optimized insulation deterioration index continuously exceeds the limit and compensation correction is ineffective, activating a field intensity map reconstruction instruction; obtaining compensated spatial insulation field intensity distribution characteristic information at the current time; performing insulation field intensity spatial interpolation calculation by fusing a device structure topology map; generating a new device insulation field intensity map with three-dimensional coordinate attributes.

7. The on-line monitoring system for insulation condition of power equipment according to claim 1, characterized in that, The system further includes an abnormality disposal decision module configured to generate an insulation state maintenance strategy based on the reconstructed device insulation field intensity map, including: identifying field intensity weak area coordinates in the new device insulation field intensity map; associating an operation and maintenance procedure library of the power device to match a disposal plan template; combining a real-time load rate of the device to generate a disposal priority sequence; outputting an insulation state maintenance strategy containing disposal location coordinates, disposal methods, and execution timing.

Images